1: \documentclass[12pt,preprint]{aastex}
2: \usepackage{natbib}
3: \usepackage{amsmath}
4: %% manuscript produces a one-column, double-spaced document:
5:
6: %\documentclass[manuscript]{aastex}
7:
8: %% preprint2 produces a double-column, single-spaced document:
9:
10: % \documentclass[preprint2]{aastex}
11:
12: %% If you wish, you may supply running head information, although
13: %% this information may be modified by the editorial offices.
14: %% The left head contains a list of authors,
15: %% usually a maximum of three (otherwise use et al.). The right
16: %% head is a modified title of up to roughly 44 characters. Running heads
17: %% will not print in the manuscript style.
18:
19: \shorttitle{A Radio-Loud NLS1}
20: \shortauthors{Oshlack, Webster, Whiting}
21:
22: \begin{document}
23:
24: \title{A Very Radio-Loud Narrow-Line Seyfert 1: PKS 2004-447}
25:
26: \author{A. Y. K. N. Oshlack,}
27: \affil{School of Physics, University of Melbourne, Parkville, Victoria, Australia, 3010}
28: \email{aoshlack@physics.unimelb.edu.au}
29:
30:
31: \author{R. L. Webster}
32: \affil{School of Physics, University of Melbourne, Parkville, Victoria, Australia, 3010}
33: \email{rwebster@physics.unimelb.edu.au}
34:
35: \author{and M. T. Whiting}
36: \affil{School of Physics, University of Melbourne, Parkville, Victoria, Australia, 3010}
37: \email{mwhiting@physics.unimelb.edu.au}
38:
39: \begin{abstract}
40:
41: We have discovered a very radio-loud Narrow-Line Seyfert 1 candidate:
42: PKS 2004-447. This Seyfert is consistent with the formal definition
43: for NLS1s, although it does not have quite the same spectral features
44: as some typical members of this subclass. Only ROSAT survey data is available at X-ray wavelengths, so it has not been possible to
45: compare this source with other NLS1s at these wavelengths. A full
46: comparison of this source with other members of the subclass will
47: improve our physical understanding of NLS1s. In addition,
48: using standard calculations, we estimate the central black hole to have
49: a mass of $\sim 5 \times 10^6 M_{\odot}$. This does not agree with
50: predictions in the literature, that radio-loud
51: AGN host very massive black holes.
52:
53: \end{abstract}
54:
55: \keywords{galaxies:active --- galaxies:Seyfert ---
56: quasars:individual(PKS 2004-447)}
57:
58:
59:
60: \section{Introduction}
61: The radio-loud source PKS 2004-447, $z=0.24$, was identified as a candidate
62: Narrow-Line Seyfert 1 (NLS1) from low
63: resolution optical spectra of a subsample of quasars identified in the Parkes
64: Half-Jansky Flat-Spectrum Sample \citep{dri97}(hereafter PHFS). The
65: PHFS is designed to efficiently select radio-loud AGN
66: by looking for flat or inverted radio spectra over the range of 2.7
67: and 5.0 GHz. It contains 323 radio bright ($>0.5$Jy at 2.7 GHz),
68: generally compact radio sources. Higher resolution spectroscopy of
69: this object was obtained giving more accurate velocity widths of the
70: emission lines indicating this may be an NLS1 but more accurate
71: spectroscopy is still needed to be able to clearly separate the broad
72: and narrow components of H$\beta$. In any case this object has very
73: interesting implications to the study of AGN.
74:
75: NLS1s are generally considered to be an extreme but common subclass of
76: AGN. They are defined by their optical emission line
77: properties, such that the H$\beta$ line is both strong (with a flux ratio
78: [\ion{O}{3}]/H$\beta<3$, similar to Seyfert 1 galaxies) and narrow (the
79: H$\beta_{FWHM}<2000$km s$^{-1}$) \citep{ost85}. These properties
80: are correlated with strong \ion{Fe}{2} emission and a strong soft
81: X-ray excess,
82: amongst other properties, for most objects in the subclass.
83: NLS1s do not appear to form a distinct class
84: but are instead connected to the ``standard'' broad-line Seyfert
85: population through a continuum of properties. \citet{bor92}
86: have shown that NLS1s cluster at one end of the region defined
87: by the first principal component described in their study of the
88: optical spectra of 87 low-redshift BQS quasars. The first
89: principal component is
90: thought to describe the major physical property in quasar
91: structure that is responsible for spectral differences between
92: AGN, independent of orientation. \citet{bor92}
93: suggest that the physical parameter driving this eigenvector is
94: $\dot{M} / \dot{M}_{Edd}$, where $\dot{M}$ is the rate of mass accretion onto the central massive object. NLS1s are thought to be accreting
95: at a rate closer to the Eddington limit ($\dot{M}_{Edd}$) \citep{bor92,lao97}.
96: NLS1s are generally radio-quiet objects, with only three
97: previously identified radio-loud objects, PKS 0558-504 \citep{rem86}, RGB
98: J0044+193 \citep{sie99} and J0134.2-4258 \citep{gru00}.
99: \citet{bor92} find that
100: radio-loud QSOs and the NLS1s (with a strong, soft X-ray excess) lie at
101: opposite ends of the primary eigenvector.
102:
103: Although the criteria for NLS1s are well defined, it is unclear whether these
104: phenomenological attributes reflect a single underlying physical mechanism.
105: Thus the discovery of a very radio-loud NLS1 may indicate that the
106: observational definition of NLS1s requires refinement. Alternatively,
107: radio-loud NLS1s may
108: provide a more stringent test of the models of NLS1s.
109: Three important consequences of the identification of a radio-loud NLS1
110: will be considered. Firstly, are the radio-loud NLS1s the same
111: class of objects as most others in the class. Secondly, are radio-loud
112: NLS1s consistent with
113: any of the popular models for NLS1s and thirdly, do radio-loud quasars
114: require large mass black
115: holes. This object challenges that assertion as is further discussed in
116: section~4
117:
118: In section 2, we present the observational data on PKS 2004-477.
119: In section 3, this object is compared to the three other radio-loud NLS1s
120: which have been identified. The central black hole mass is determined
121: using standard techniques in section 4, and possible models for NLS1s
122: are discussed in section
123: 5. Finally our conclusions are presented in section 6.
124:
125:
126:
127: \section{Observational Data}
128:
129: \subsection{Optical Spectrum}
130:
131: PKS 2004-447 was first identified as a NLS1 candidate from a low
132: resolution spectrum obtained using the RGO/FORS spectrograph at the AAT in
133: 1984 and published in \citet{dri97}.
134: A higher resolution confirmation spectrum
135: was taken using the double beam spectrograph (DBS) on the ANU 2.3m
136: telescope, 1st August 2000, and is shown in Fig~1. The conditions were
137: not photometric. The spectrum was reduced using standard
138: procedures in the IRAF. The spectrum has a resolution
139: of$\sim 2.2$\AA.
140:
141: \begin{figure}
142: \plotone{f1.eps}
143: \caption{Rest frame, medium resolution spectrum of PKS~2004-447 from the
144: DBS on the 2.3m telescope at the Siding Springs Observatory, showing the
145: H$\beta$ and [\ion{O}{3}] region (top) and the best-fit rescaled \ion{Fe}{2} spectrum shown
146: below it.}
147: \end{figure}
148:
149: \begin{figure}
150: \plotone{f2.eps}
151: \caption{Low resolution spectrum of PKS~2004-447 from RGO spectrograph
152: on the AAT.}
153: \end{figure}
154:
155: Typically, in NLS1s, \ion{Fe}{2} emission contaminates the spectrum making
156: accurate measurements of H$\beta$ and [\ion{O}{3}] difficult. In order to
157: account for this and to get an estimate on the strength of the \ion{Fe}{2}
158: emission in this object we employed the \ion{Fe}{2} subtraction method
159: introduced by \citet{bor92} and now commonly used. We use a template
160: \ion{Fe}{2} spectrum taken from the prototype strong Fe emitter I Zw 1 and
161: scale it and shift it until the width and intensity match those seen
162: in our object. We did this over the region 5050-5500\AA~ using a
163: $\chi ^2$-minimisation to get the best value
164: for the scaling. We find that this spectral region contains very
165: little \ion{Fe}{2} emission with an
166: equivalent width EW$< 10$\AA~for the whole \ion{Fe}{2} complex in the region
167: 5050-5450 \AA. Compared to values measured in \citet{bor92}, this
168: is extremely low (although we are fitting a slightly different
169: region of the \ion{Fe}{2} spectrum).
170: We measured the line widths and fluxes using Lorentzian
171: fits to the emission lines and the results are summarised in Table 1.
172: The width of $H\beta _{FWHM}=1447$km s$^{-1}$and the flux ratio
173: [\ion{O}{3}]/H$\beta= 1.6$,
174: fit the criteria for
175: classification as a NLS1. The DBS spectrum does not extend to
176: wavelengths of the \ion{Fe}{2}($\lambda$4570) lines. However the previous low
177: resolution spectrum indicates \ion{Fe}{2} emission in the region 4435-4700
178: \AA, blueward of the $H\beta$
179: line (Fig 2). The low resolution spectrum also shows strong H$\alpha$
180: and evidence for H$\gamma$ emission and it also indicates that the
181: stregth of the H$\beta$ emission, relative to [\ion{O}{3}] has varied
182: between the two epochs.
183:
184:
185: PKS 2004-447 is also optically variable. The magnitude of
186: this object measured from the COSMOS/UKST Southern Sky Catalogue is
187: $B_{J}=18.1$, while more recent photometry obtained
188: on the ANU 2.3m telescope gave $B=19.5$ \citep{fra00}, indicating a
189: drop in flux by a factor of $\approx 4$ over an interval of
190: several years. Simultaneous optical/IR photometry are shown in
191: Fig~3. The continuum is very red and there is no
192: evidence of a big blue bump.
193: The absolute magnitude is between -19.0 and -21.2, for $q_{0} =0.5$ and
194: $H_0 =100$, clearly placing this in the Seyfert luminosity class.
195:
196: \begin{table}
197: \begin{center}
198: \caption{ Equivalent widths of the emission lines\tablenotemark{a}}
199: \begin{tabular}{|p{3.5cm}||p{2.5cm}|p{2.5cm}|p{2.5cm}|}
200: \hline & H$\beta$ (4861) & \ion{O}{3} (4959) & \ion{O}{3} (5007)\\
201: \hline
202: \hline FWHM(km s$^{-1}$) & 1447 & 951 & 754 \\
203: \hline flux relative to H$\beta$ & 1.0&0.61 &1.61 \\
204: \hline EW (\AA)& 31.6&19.6 &50.6 \\
205: \hline Offset (km s$^{-1}$)& $+28$&$-78$&0\\
206: \tableline
207: \end{tabular}
208: \tablenotetext{a}{Measured from the spectrum in Fig~1; details are described in the text.}
209: \end{center}
210: \end{table}
211:
212:
213: \subsection{Radio Emission}
214: Under the selection criteria, the PHFS objects are detected
215: above 0.5 Jy at 2.7GHz and with a spectral slope $\alpha < 0.5$
216: (F$_{\nu}\propto \nu^{-\alpha}$) taken from
217: non-simultaneous observations at 2.7GHz and 5.0 GHz. Using this method PKS
218: 2004-447 had a 2.7 GHz flux of 0.81 Jy and a spectral index $\alpha_{r} =
219: 0.36$. Subsequent simultaneous observations of this source were taken
220: using the ATCA (23rd November 1995) and the radio flux had
221: varied since the Parkes observations,
222: demonstrating long term variability at these
223: frequencies as well. The radio spectrum is shown in Fig~3. A powerlaw
224: fit to the simultaneous data gives a spectral index $\alpha_{r} = 0.67$;
225: thus PKS 2004-447 is actually a steep spectrum source. Calculating
226: $R$, where $R$ is the ratio of radio to optical flux ($f_{4.85GHz}/f_{B}$), \citep{kel89}, gives
227: values in the range $1710 < R < 6320$ depending on value used for the
228: optical magnitude. Obviously, this source is very radio-loud.
229:
230: \begin{figure}
231: \plotone{f3.eps}
232: \caption{Spectral energy distribution of PKS 2004-447. Radio data from
233: contemporaneous observations at the ATCA (circles, unpublished data). Optical data from
234: quasi-simultaneous observation at Siding Spring \citep[squares]{fra00}. X-ray data from ROSAT
235: All Sky Survey \citep[line]{sie98}.}
236: \end{figure}
237:
238: The radio image is
239: unresolved in the ATCA observations.
240: The visibility of PKS 2004-477 was measured using the Parkes-Tidbinbilla Interferometer on a 275 km baseline at 2.3 GHz \citep{dun93}. The average
241: of two measurements gave a visibility of 0.680 corresponding to an
242: angular size of $\sim 0.036 \arcsec$ or $\sim 85 pc$. Since this source size
243: is derived from a 2-element interferometer, it is measured along one baseline.
244:
245: There are two previous studies of the radio properties of NLS1s. The
246: most recent and well defined is that of \citet{mor00}, who obtained high
247: resolution observations of 24 NLS1s at 20cm and 3.6cm using the VLA.
248: All but one of the sources were detected, with 20cm fluxes up to $10^{25} W\,
249: Hz^{-1}$, which is higher, on average, than classical Seyferts.
250: Moran found most NLS1s unresolved at $\sim 1 \arcsec$, with
251: generally quite steep ($\alpha \approx 1.1-1.2$) radio spectra, though the
252: spectral index of PKS 2004-477 is consistent with the flattest
253: values. In other words these could be considered as low luminosity
254: compact steep spectrum (CSS) sources.
255: There is also a few cases of variability at 20cm, again,
256: as seen in PKS 2004-477. Thus PKS 2004-477 is compact, steep-spectrum
257: and variable in the radio, which are similar properties to other NLS1s
258: \citet{mor00}. This is further discussed in section 5.
259:
260:
261:
262: \subsection{X-ray Emission}
263: NLS1s have unusual X-ray properties. They generally display a
264: large soft X-ray excess, a steep hard X-ray spectrum and large
265: amplitude X-ray variability over short and long timescales.
266: \citet{sie98} investigated the X-ray properties of the PHFS using the
267: ROSAT All Sky Survey. PKS 2004-447 was detected with a total
268: flux in the 0.1-2.4 keV range of $0.427 \pm 0.218 \times 10^{-12}$ erg
269: cm$^{-2}$ s$^{-1}$ corresponding to $L_X = 2.9 \times 10^{43} $erg
270: s$^{-1}$. We find an optical--to--X-ray slope of
271: $\alpha_{ox}=-1.2$. This is consistent with values found by
272: \citet{xu99} for a ROSAT selected sample of NLS1s.
273:
274: \section{Previous Detections of Radio-Loud NLS1s}
275:
276: Most NLS1s are radio-quiet and only three radio-loud NLS1s are
277: known \citep{sie99}.
278: PKS 0558-504 was optically identified on the basis of X-ray positions
279: from the High Energy Astronomical Observatory \citep{rem86}.
280: It was noted as not only having very narrow hydrogen emission lines,
281: but also strong \ion{Fe}{2} emission. It has a radio flux of 113 mJy at 4.85
282: GHz, an optical
283: magnitude $m_B = 14.97$ and a redshift $z=0.137$. This gives $R \approx 27$
284: \citep{sie99}. This source exhibits strong X-ray variability
285: on medium (months) and short (days, hours) time scales \citep{gli00}
286: and an X-ray flare indicating relativistic beaming \citep{rem91}.
287:
288: \citet{sie99} identified RGB J0044+193 as a radio-loud NLS1.
289: RGB J0044+193 has only a moderate radio flux, but is
290: calculated to have a
291: radio-to-optical flux ratio $R \approx 31$.
292: It was identified as a radio-loud X-ray
293: source with a redshift of $0.181$ in a cross-correlation of
294: the ROSAT All-Sky Survey and the
295: Green Bank 5GHz radio survey \citep{lau98}. It has the bluest optical
296: spectrum of
297: any NLS1 observed to date. \citet{sie99} measure an optical continuum
298: slope of $\alpha = 1.3$ redward of 5000\AA\ and a slope of $-3.1$ blueward
299: of 5000\AA\, where $f_{\nu} \propto \nu^{-\alpha}$.
300: This source was detected in the 87GB
301: survey with $f_{4.85GHz}=24$ mJy. A high resolution follow-up using
302: the VLA at 4.85 GHz gave a flux of only 7 mJy \citep{lau97},
303: indicating that this
304: source is extended or variable. The latter is supported by the fact
305: that the radio source is unresolved on the VLA map and it is not
306: detected at 1.4 GHz in the NRAO/VLA Sky Survey, which is sensitive down
307: to 2.5 mJy. This suggests that the classification of the source as radio-loud
308: is uncertain. In all other respects, RGB J0044+193 is indistinguishable from
309: radio-quiet NLS1s.
310:
311: RX J0134.2-4258 was discovered by \citet{gru00} to have the steepest
312: soft X-ray spectrum observed during the ROSAT All-Sky Survey. It has an
313: optical magnitude of $m_V =16.2$ and $z=0.237$. It was detected in the
314: PMN survey \citep{wri94} with a flux of 0.055 Jy at 4.85 GHz,
315: and was subsequently re-observed at the VLA to have
316: a flux of 0.025 Jy at 8.4 GHz. This gives a ratio of $R=71$, and a
317: radio spectral index of $\alpha = 1.4$ for non-simultaneous observations.
318: The source is highly variable in the X-ray, and there is evidence for
319: variability at other wavelengths. The \ion{Fe}{2} emission is strong, unlike
320: that of PKS 2004-477.
321:
322: \section{Black Hole Mass}
323:
324: There seems to be a growing consensus that a more
325: massive black hole is needed to produce a radio-loud quasar
326: \citep{lao00,kas00,mcl00,pet00}. PKS~2004-447 appears to
327: directly conflict with this
328: statement. Calculations of the black hole mass rely on the assumption
329: that the dynamics of the BLR gas are dominated by the central black
330: hole. To calculate the black hole mass for this object we use the
331: results of \citet{kas00}, which are derived from a
332: reverberation study of 17 quasars.
333: \citet{kas00} empirically determine linear relationship between
334: $R_{BLR}$ and the luminosity ($\lambda L_{\lambda}$) which we use to estimate
335: the radius of the BLR-emitting gas. From equation~5 in \citet{kas00}.
336: \begin{equation}
337: R_{BLR} = (32.9)\left[\frac{\lambda L_{\lambda}(5100
338: {\text \AA})}{10^{44} {\text erg~s}^{-1}}\right]^{0.700} \text{lt-days}
339: \end{equation}
340: We use the luminosity taken from a linear interpolation between
341: photometric data points \citep{fra00} at the rest wavelength of 5100\AA.
342: Following \citet{kas00}, the mass of the black hole is given by
343: $M_{BH}=rv^{2}G^{-1}$. To determine $v$, the velocity, we correct
344: $v_{FWHM}$ of the H$\beta$ emission line by a factor of $\sqrt{3}/2$ to account for velocities in
345: three dimensions. The mass is then
346: \begin{equation}
347: M = 1.464 \times 10^{5} \left(\frac{R_{BLR}}{\text{
348: lt-days}}\right)\left(\frac{v_{FWHM}}{10^{3}{\rm km~s}^{-1}}\right)^{2}
349: M_{\odot}
350: \end{equation}
351: Using the cosmology $q_{0} =0.5$ and $H_{0}=100$km s$^{-1}$Mpc$^{-1}$,
352: this gives us a value for the central black
353: hole mass of $5.4 \times 10^6 M_{\odot}$. This
354: mass is two orders of magnitude lower than those obtained by other
355: authors for radio-loud quasars. In
356: a study of the black hole masses of quasars from the \citet{bor92}
357: sample, it was found that all quasars with $M_{BH}<3 \times 10^8
358: M_{\odot}$ are radio-quiet \citep{lao00}. The data from this sample is shown
359: in Fig~4, with the mass of PKS~2004-447 added. It can be seen that
360: PKS~2004-447 is
361: quite different than the rest of the sample. We note that
362: \citet{lao00} assumes a flatter relationship than \citet{kas00}, but
363: estimates that this difference results in a discrepancy of $<50\%$ in
364: the mass estimates. This corresponds to a shift of $<0.3$ in
365: log$M_{BH}$, which does not qualitatively affect the result in
366: figure~4. Similarly
367: \citet{mcl00}, in their study of radio-loud and radio-quiet objects,
368: came to the conclusion that a radio-loud quasar requires a mass
369: $M_{BH}>6 \times 10^8 M_{\odot}$. This again,is more than two orders
370: of magnitude
371: greater than the mass we derive for PKS~2004-447.
372:
373: We speculate that
374: the reason our object has a black hole mass so much smaller than
375: those seen in previous studies is not entirely due to it being an
376: unusual object, but more likely due to the selection criteria in the
377: previous samples. All previous studies have used optical
378: selection to compile their sample. This selection technique
379: preferentially selects radio-loud
380: quasars which are blue with very broad emission lines. However samples of
381: radio-selected quasars such as the PHFS show that the general radio-loud
382: population can be quite different \citep{web95,fra00}. In these
383: samples it may be
384: quite common to have low mass black holes producing large radio-jets
385: and this issue will be addressed in a future paper.
386:
387: \begin{figure}
388: \plotone{f4.eps}
389: \caption{Radio loudness vs. black hole mass, showing PKS 2004-447 in
390: relation to data from figure~2 of \citep{lao00}. The \citet{lao00}
391: data points are from the \citet{bor92} sample of 87 low redshift
392: quasars. The circled squares were proposed by \citet{fal96a,fal96b} to
393: be beamed intrinsically radio-quiet quasars. PKS~2004-447 has been
394: placed on the diagram at its lowest value for $R$ quoted in text.}
395: \end{figure}
396:
397:
398: \section{Discussion of Physical models}
399:
400: \subsection{Classification}
401:
402: Since the subclass of NLS1s are defined by phenomenological
403: measurements, rather than a specific physical model, it
404: is possible that more than one physical mechanism will produce
405: the defining parameters. First we should consider whether
406: PKS~2004-447 should be defined as an NLS1. Certainly it meets
407: the formal definition of \citet{ost85}. In the X-ray, the
408: flux ratios are consistent with other NLS1s, but pointed observations
409: are required to obtain spectral information. In the optical,
410: the \ion{Fe}{2} emission is much weaker than other NLS1s. Apart from the
411: obvious strength of the radio
412: flux, the size of the radio source, and its spectral index are
413: consistent with other radio measurements \citep{mor00}.
414:
415: The object's radio properties fit with the definition of a Compact
416: Steep Spectrum radio source (CSS) according to the definition from \citet{fan95}
417: of: $P_{tot} > 10^{26}$W Hz$^{-1}$; a size
418: smaller or comparable to the optical galaxy scale; and the spectral
419: index $\alpha_{r} >0.5$. The CSS sources tend to have spectra with
420: narrow emission lines typical of radio-galaxies.
421: The broad component
422: of H$\beta$ seems to be extremely weak or absent
423: \citep{mor97}. \citet{gel94} find an average value [\ion{O}{3}]/H$\beta \sim
424: 5$ indicating that the H$\beta$ flux is too weak to be an NLS1. In the study
425: of \citet{gel94} all CSS sources with measurable H$\beta$ are of
426: quasar luminosity. So PKS 2004-447 is unusual in both categories: firstly it has
427: an optical spectrum which is unusual for CSS source; and secondly has
428: radio power unusual for NLS1s.
429:
430: The source is clearly in the Seyfert luminosity range, as are
431: a significant fraction of the sources in PHFS.
432: The differences in optical spectral properties, (such as the weakness of
433: the \ion{Fe}{2} emission) between
434: PKS 2004-477 and NLS1s in general raise the possibility
435: that the NLS1-defining characteristics of this source are
436: due to a different physical mechanism
437: from other objects in the subclass. This possibility may be
438: resolved when an X-ray spectrum is obtained.
439:
440: \subsection{Physical Models}
441:
442: Finding unusual objects in a particular class is a useful
443: method for distinguishing between different possible physical models.
444: Although there have been a variety of
445: physical models suggested for NLS1s \citep[see for a summary]{tan99}
446: it seems that the model with the strongest observational support
447: is the high accretion rate, low black hole mass model.
448: This model suggests that NLS1s are objects where the black
449: hole is accreting at a rate closer to the Eddington limit compared to
450: broad-line Seyferts. Since there is a higher rate of accretion,
451: a given luminosity corresponds to a black hole with a smaller mass.
452: If the size of the BLR is determined by the bolometric luminosity, then,
453: at a given radius, the broad line clouds will have slower velocities
454: and the emission lines will be narrower.
455: This model was developed to explain the soft X-ray excess and the
456: steep X-ray powerlaw, by analogy with the X-ray spectra of galactic
457: black hole candidates accreting in their ``high''state \citep{pou95}.
458: For high accretion rates, the soft thermal emission from the
459: disk becomes energetically dominant, producing the soft X-ray excess,
460: and the X-ray powerlaw spectrum can be steepened as a result of Compton
461: cooling of the electrons by the soft X-ray photons.
462:
463: An alternative model for the narrow Hydrogen lines suggests we are
464: viewing the source near the axis of the accretion disk. If the
465: velocity width of the Hydrogen emission lines is dominated by the
466: rotation of the accretion disk, then they will appear narrow.
467: The radio emission in PKS is moderately steep spectrum, which
468: weakens the argument that we observe this source near the axis.
469: Indeed it also weakens arguments which suggest that the radio
470: emission is strongly boosted by beaming. However the radio
471: source is very compact, with no evidence at any frequency for
472: double-lobed emission.
473:
474: \section{Conclusion}
475:
476: PKS 2004-447, detected in the PHFS, is an unusual source. To summarize the major points regarding PKS~2004-447:
477: \begin{enumerate}
478: \item It has H$\beta$ width and strength consistent with being
479: classified as an NLS1 as defined by \citet{ost85}.
480: \item Its radio properties are very unusual for a Seyfert Galaxy
481: having the following characteristics:
482: very strong radio flux, 0.81 Jy at 2.7 GHz;
483: very radio-loud, $R>1700$;
484: steep radio spectral index $\alpha_{r}=0.67$;
485: compact radio source;
486: some evidence of long term variability.
487: These properties are consistent with it being classified as a
488: Compact Steep Spectrum source.
489: \item It has been detected in the RASS with a flux $0.427 \times
490: 10^{-12}$ erg~cm$^{-2}$~s$^{-1}$ but no spectral information is available.
491: \item Following the procedure set out by \citet{kas00}, we calculate
492: the black hole mass to be $5.4 \times 10^{6} M_{\odot}$. This mass is
493: more than 2 orders of magnitude lower than those seen previously for
494: radio-loud AGN and challenges previous results that a large black hole
495: mass is needed to produce radio-loud AGN.
496: \end{enumerate}
497: Additional observations of the X-ray
498: spectrum will further constrain models and may provide evidence that
499: the black hole is accreting at a rate closer to the Eddington
500: limit. This would be an indication that the underlying physical mechanism
501: in this object is similar to those in other NLS1s, and that
502: the radio power of the object is actually connected to a different
503: parameter.
504:
505: We wish to thank Dirk Grupe for providing us with the
506: \ion{Fe}{2} template used in the analysis.
507:
508: \begin{thebibliography}{}
509: %
510: \bibitem[Boroson and Green (1992)]{bor92} Boroson, T.\ A. and Green,
511: R.\ F. 1992 \apjs, 80, 109
512: %
513: \bibitem[Drinkwater et al.(1997)]{dri97} Drinkwater, M.\ J.,
514: Webster, R.\ L., Francis, P.\ J., Condon, J.\ J., Ellison, S.\ L.,
515: Jauncey, D.\ L., Lovell, J., Peterson, B.\ A., \&
516: Savage, A. 1997, \mnras, 284, 85
517: %
518: \bibitem[Duncan et al. (1993)]{dun93} Duncan, R. A., White, G. L., Wark, R., Reynolds, J. E., Jauncey, D. L., Norris, R. P., Taaffe, L. and Savage, A.,
519: 1993, PASA, 10, 310
520: %
521: \bibitem[Falcke, Patnaik, \& Sherwood(1996)]{fal96a} Falcke,
522: H., Patnaik, A.\ R., \& Sherwood, W.\ 1996a, \apjl, 473, L13
523: %
524: \bibitem[Falcke, Sherwood, \& Patnaik(1996)]{fal96b} Falcke,
525: H., Sherwood, W., \& Patnaik, A.\ R.\ 1996b, \apj, 471, 106
526: %
527: \bibitem[Fanti et al.(1995)]{fan95} Fanti, C., Fanti, R.,
528: Dallacasa, D., Schilizzi, R.\ T., Spencer, R.\ E., \& Stanghellini, C.\
529: 1995, \aap, 302, 317
530: %
531: \bibitem[Francis, Whiting, \& Webster(2000)]{fra00} Francis,
532: P.\ J., Whiting, M.\ T., \& Webster, R.\ L.\ 2000, Publications of the
533: Astronomical Society of Australia, 17, 56
534: %
535: \bibitem[Gelderman and Whittle (1994)]{gel94} Gelderman, R., \&
536: Whittle, M., 1994, \apjs, 91, 491
537: %
538: \bibitem[Gliozzi et al.(2000)]{gli00} Gliozzi, M., Boller, Th., Brinkmann, W.
539: and Brandt, W. N., 2000, \aap, 356, 17
540: %
541: \bibitem[Grupe, Leighly, Thomas, \&
542: Laurent-Muehleisen(2000)]{gru00} Grupe, D., Leighly, K.\ M.,
543: Thomas, H.\ -., \& Laurent-Muehleisen, S.\ A.\ 2000, \aap, 356, 11
544: %
545: \bibitem[Kaspi et al.(2000)]{kas00} Kaspi, S., Smith, P.\ S.,
546: Netzer, H., Maoz, D., Jannuzi, B.\ T., \& Giveon, U.\ 2000, \apj, 533, 631
547: %
548: \bibitem[Kellerman et al.(1989)]{kel89} Kellerman, K. I., Sramek, R.,
549: Schmidt, M., Shaffer, D. B., Green, R., 1989, \aj, 98, 1195
550: %
551: \bibitem[Laor et al.(1997)]{lao97} Laor, A., Fiore, F.,
552: Elvis, M., Wilkes, B.\ J., \& McDowell, J.\ C.\ 1997, \apj, 477, 93
553: %
554: \bibitem[Laor(2000)]{lao00} Laor, A.\ 2000, \apjl, 543, L111
555: %
556: \bibitem[Laurent-Muehleisen et al.(1997)]{lau97}
557: Laurent-Muehleisen, S.\ A., Kollgaard, R.\ I., Ryan, P.\ J., Feigelson, E.\
558: D., Brinkmann, W., \& Siebert, J.\ 1997, \aaps, 122, 235
559: %
560: \bibitem[Laurent-Muehleisen et al.(1998)]{lau98}
561: Laurent-Muehleisen, S.\ A., Kollgaard, R.\ I., Ciardullo, R., Feigelson,
562: E.\ D., Brinkmann, W., \& Siebert, J.\ 1998, \apjs, 118, 127
563: %
564: \bibitem[Moran (2000)]{mor00} Moran, E, C., 2000, NewAR, 44, 527
565: %
566: \bibitem[Morganti (1997)]{mor97} Morganti, R., Tadhunter, C.\ N.,
567: Dickson, R., Shaw, M., \aap, 326, 130
568: %
569: \bibitem[McLure and Dunlop (2000)]{mcl00} McLure, R.\ J. \& Dunlop, J.\
570: S., astro-ph/0009406.
571: %
572: \bibitem[Osterbrock and Pogge (1985)]{ost85} Osterbrock, D. E. and
573: Pogge, R. W., 1985, \apj, 297, 166
574: %
575: \bibitem[Peterson et al.(2000)]{pet00} Peterson, B.\ M.\ et
576: al.\ 2000, \apj, 542, 161
577: %
578: \bibitem[Pounds et al.(1995)]{pou95} Pounds, K. A., Done, C. and Osborne, J. P.,
579: 1995 \mnras, 277, L5
580: %
581: \bibitem[Remillard et al.(1986)]{rem86} Remillard, R. A., Bradt, H. V., Buckley, D. A. H., Roberts, W., Schwartz., D. A., Tuohy, I. R \& Wood, K. 1986, \apj,
582: 301, 742
583: %
584: \bibitem[Remillard et al.(1991)]{rem91} Remillard, R. A., Grossan, B.,
585: Bradt, H. V., Ohashi, T. \& Hayashida, K., 1991 \nat, 350, 589
586: %
587: \bibitem[Siebert et al.(1998)]{sie98} Siebert, J., Brinkman, W., Drinkwater, M. J.,
588: Yuan, W., Francis, P. J., Peterson, B. A. \& Webster, R. L. 1998, \mnras,
589: 301, 261
590: %
591: \bibitem[Siebert et al.(1999)]{sie99} Siebert, J., Leighly, K. M.,
592: Laurent-Muehleisen, S. A., Brinkman, W., Boller, \& Matsuoka, M., 1999,
593: \aap, 348, 678
594: %
595: \bibitem[Taniguchi et al.(1999)]{tan99} Taniguchi, Y., Murayama, T., \&
596: Nagao, T., 1999, astro-ph/9910036
597: %
598: \bibitem[Webster et al.(1995)]{web95} Webster, R.\ L.,
599: Francis, P.\ J., Peterson, B.\ A., Drinkwater, M.\ J., \& Masci, F.\ J.\
600: 1995, \nat, 375, 469
601: %
602: \bibitem[Wright et al.(1994)]{wri94} Wright, A. E., Griffith, M. R.,
603: Burke, B. F. \& Ekers, R. D., 1994, \apjs, 91, 111
604: %
605: \bibitem[Xu, Wei \& Hu (1999)]{xu99} Xu, D.\ W., Wei, J.\ Y. \& Hu, J.\ Y.,
606: 1999, \apj, 517, 622
607: %
608: \end{thebibliography}
609:
610: \end{document}
611:
612:
613: